Lucene : Where to use exactly

[sandeep lokhande]

Welcome Michael McCandless, Erik Hatcher, and Otis Gospodnetic,

I am building a application for Report generation for my some other application,
m using Jasper reports for it.
how i can use Lucene with my system for report creation.
Thank you.

and Welcome Once Again.

[Lester Burnham]

What functionality are you looking for that JasperReports doesn't provide? It's a bit hard to see where Lucene would fit in with report generation (you know that it's a library for indexing and searching text, right?). Unless you're thinking of using a Lucene index as an alternative data source instead of a DB, or wherever your report data is normally coming from. But that would be an unusual design.

[Kaustubh G Sharma]

Hi I am working with jasper reports and face some problem like it don't provide table and if you find my posts mostly related to jasper report only so is this Lucene overcame those problems if YES... how?

[Jaikiran Pai]

so is this Lucene overcame those problems if YES... how?

Lucene and Jasper Reports are two different products. Lucene is full-text search library whereas Jasper Reports is a reporting tool. For more details on what Lucene is, see the reply in your other thread in this forum.

[sandeep lokhande]

I Used to depend upon My DB for searching Data and Querying for it.

What are the advantages in Lucene other than my relational DB.

[Lester Burnham]

I think you have a rather incorrect impression of what Lucene is and does. The OtherOpenSourceProjectsFaq links to several introductory articles that should get you started.

[Aneesh Vijendran]

sandeep lokhande wrote:I Used to depend upon My DB for searching Data and Querying for it.

What are the advantages in Lucene other than my relational DB.

Briefly:

I used Lucene a few years back to power one of our Document Management Systems.

Though all document details, such as meta data, etc are stored in the DB we used Lucene for search.

Reason:


Users mostly need to search from within the documents and opening a document and parsing it each time when a user request a keyword, would be suicidal.
So while creating/uploading/updating a document in the DMS we index using Lucene (which updates the Lucene index). We extract teh text/body from docuemnts and pass it to Lucene to add it to the index, aslon with other metadatas.
For a few million documents the Lucene Index grew a lot of Giga Bytes of data, and surprisingly our search from Lucene returned in less than 3-4 seconds.

Moreover it had lot of Search features which none of the DB's at that time(and still) doesn't support, such as:

ranked searching, phrase queries, wildcard queries, proximity queries, range queries, date, Sounds like (I can't find that in the recent versions somehow).

An interesting problem we faced was, as the Index grew so big. Due to some reason once the index got corrupt. We had to write a tool for re-indexing the whole repository. But this problem kept on coming back and re-indexing took ages.

A solution we found was to split/merge the index.

I would say - Lucene is a very carefully thought & brilliantly designed search Api.

PS- don't compare your DB search with Lucene. Both is entirely different and used in totally different contexts. Hope this clears all confusions.Please feel free to ask and I'll be happy to answer any queries regarding Lucene, as I had created a "lightning Search" using this Api.

Best Regards
Aneesh
http://uk.linkedin.com/in/aneeshvijendran

[sandeep lokhande]

Aneesh Vijendran wrote:

sandeep lokhande wrote:I Used to depend upon My DB for searching Data and Querying for it.

What are the advantages in Lucene other than my relational DB.

Briefly:

I used Lucene a few years back to power one of our Document Management Systems.

Though all document details, such as meta data, etc are stored in the DB we used Lucene for search.

Reason:


Users mostly need to search from within the documents and opening a document and parsing it each time when a user request a keyword, would be suicidal.
So while creating/uploading/updating a document in the DMS we index using Lucene (which updates the Lucene index). We extract teh text/body from docuemnts and pass it to Lucene to add it to the index, aslon with other metadatas.
For a few million documents the Lucene Index grew a lot of Giga Bytes of data, and surprisingly our search from Lucene returned in less than 3-4 seconds.

Moreover it had lot of Search features which none of the DB's at that time(and still) doesn't support, such as:

ranked searching, phrase queries, wildcard queries, proximity queries, range queries, date, Sounds like (I can't find that in the recent versions somehow).

An interesting problem we faced was, as the Index grew so big. Due to some reason once the index got corrupt. We had to write a tool for re-indexing the whole repository. But this problem kept on coming back and re-indexing took ages.

A solution we found was to split/merge the index.

I would say - Lucene is a very carefully thought & brilliantly designed search Api.

PS- don't compare your DB search with Lucene. Both is entirely different and used in totally different contexts. Hope this clears all confusions.Please feel free to ask and I'll be happy to answer any queries regarding Lucene, as I was created "lightning Search" using this Api.

Best Regards
Aneesh
http://uk.linkedin.com/in/aneeshvijendran

Okey,

that's pretty interesting ,

I was thinking in a wrong direction actually.
I am very enthusiastic learning this.

Thank you for your explanation.